Scheduling Performance of Heavy-Tailed Data Traffic in Wireless High-Speed Shared Channels

The third generation (3G) and future wireless networks will use packet-switching technology to provide high speed data transfer aiming at a fixed broadband experience. It has been showed that Internet data traffic exhibited self-similarity and long-range dependency. Those phenomena, that may be caused by heavy-tailed packet size distributions have significant impact on the design of Internet routers; recently, such an evolution has been observed in high speed wireless channels. This article focus on radio-based schedulers and especially on High-Speed Downlink Packet Access (HSDPA) and their impact of heavy- tailed traffic. We propose two new schedulers which are designed for wireless links fed by heavy-tailed traffic and make use of the work lookahead i.e., knowledge of the remaining file size for each flow. Both proposed algorithms provide increased performance compared to traditional scheduling methods.